ATP Signaling in Brain: Release, Excitotoxicity and Potential Therapeutic Targets

Abstract

Adenosine 5′-triphosphate (ATP) is released as a genuine co-transmitter, or as a principal purinergic neurotransmitter, in an exocytotic and non-exocytotic manner. It activates ionotropic (P2X) and metabotropic (P2Y) receptors which mediate a plethora of functions in the brain. In particular, P2X7 receptor (P2X7R) are expressed in all brain cells and its activation can form a large pore allowing the passage of organic cations, the leakage of metabolites of up to 900 Da and the release of ATP itself. In turn, pannexins (Panx) are a family of proteins forming hemichannels that can release ATP. In this review, we summarize the progress in the understanding of the mechanisms of ATP release both in physiological and pathophysiological stages. We also provide data suggesting that P2X7R and pannexin 1 (Panx1) may form a large pore in cortical neurons as assessed by electrophysiology. Finally, the participation of calcium homeostasis modulator 1 is also suggested, another non-selective ion channel that can release ATP, and that could play a role in ischemic events, together with P2X7 and Panx1 during excitotoxicity by ATP.

Introduction

Adenosine 5′-triphosphate (ATP) was identified in the late 1920s and thought to be responsible for muscle contraction (Fiske and Subbarow 1929). Later, it was discovered that ATP is an essential biomolecule in the cells for energy transfer during cellular signaling cascades, and a transmitter activating membrane receptors, also known as purinoceptors, in the peripheral nervous system (Burnstock 1999). In subsequent years, evidence was obtained that ATP and purinoceptors (P2X and P2Y receptors) were also present in neuronal and non-neuronal cells of the central nervous system (Sperlágh et al. 1995). The last two decades have witnessed an explosion of interest on the research of ATP signaling that includes numerous physiological and pathophysiological functions. In this review, we summarize some of the mechanisms of ATP release, and progress in the understanding as to how this molecule causes structural and functional damage. We will also discuss recent work in our laboratory that provides information on how ATP induces in neurons a large macroscopic ion current which represents the formation of a large pore.

ATP Release in the Nervous System

ATP is a co-transmitter in both the peripheral and central nervous systems, whereby it can act as a fast excitatory neurotransmitter in many synapses (Silinsky et al. 1992; Edwards et al. 1992; Sperlágh et al. 1995). In dorsal horn neurons, for example, there are ATP excitatory receptors, raising the possibility that it plays a role in the transmission of pain (Burnstock 1999). Besides, there has been an increasing number of studies describing ATP release from brain glial cells, including astrocytes, oligodendrocytes and microglia (Guthrie et al. 1999; Queiroz et al. 1999; Jiménez et al. 2000; Cotrina et al. 2000; Ferrari et al. 1996, 1997).

ATP levels are regulated by mitochondrial oxidative phosphorylation which generates ATP from adenosine diphosphate. In general, brain cells take up glucose from the extracellular fluid and utilize it for ATP production; although glycolysis and the citric acid cycle also provide ATP, the bulk of ATP is formed by oxidative phosphorylation. However, the detailed mechanisms by which ATP is released are not well known.

ATP Release Mechanisms: the Role of P2X7 Receptors, Panx1 and CALHM1

In neurons, release of ATP can occur by conventional exocytosis and in a non-secretory manner (Al-Aqwati 1995). In fact, the P2X7 receptor (P2X7R), a member of the family of ATP-gated cation channels (Di Virgilio et al. 2009; Skaper et al. 2010), is involved in ATP release. P2X7R can form a large pore upon activation, allowing the passage of organic cations and molecules of up to 900 Da, and the leakage of metabolites including ATP (North 2002; Yan et al. 2010). Pore formation by P2X7R takes place following repetitive or prolonged stimulation with the agonist which causes sensitization of receptors, progressive increase in the ion current amplitude and a slower deactivation rate (Yan et al. 2010). In addition to neurons, glial cells also express P2X7R (North 2002; Matute et al. 2007; Verkhratsky et al. 2009). In particular, activation of P2X7R in astrocytes increases [Ca²⁺]_i and causes the release of ATP. Likewise, microglia express functional P2X7R and their activation can release ATP (Ferrari et al. 1996, 1997). On the other hand, although oligodendrocytes express P2X7R (Matute et al. 2007) it is not clear whether these cells release ATP by this mechanism.

The fact that ATP release via pore formation in P2X7R requires sustained stimulation by high extracellular concentrations of ATP has led to the proposal that this mechanism may be relevant to cell death and tissue damage in brain diseases (Yan et al. 2010). Alternatively, a role in ATP release has also been assigned to hemichannels formed by pannexin 1 (Panx1), which are large pore ion channels with broad expression in the CNS. Panx1 are permeable to molecules smaller than ~1 kD and directly mediates ATP release (Locovei et al. 2006; Iglesias et al. 2009). Panx1 is expressed in pyramidal neurons at the postsynaptic density (Ray et al. 2006; Vogt et al. 2005; Zoidl et al. 2007) and is activated following NMDA receptor stimulation, where it can contribute to bursting patterns in the hippocampus (Thompson et al. 2008). However, direct release of ATP from neurons endogenously expressing Panx1 has not yet been demonstrated. Panx1 can also be activated by high extracellular K⁺ (independently of depolarization) in cultured astrocytes (Silverman et al. 2009). The activity of Panx1 has been identified also in microglia (Orellana et al. 2012; Rigato et al. 2012). Regardless of the mechanism of Panx1 opening, there is now evidence that a key physiological role of Panx1 is to mediate ATP release. In turn, it is hypothesized that the formation of large pores in the membrane of neurons or glial cells is not due only by Panx1 or P2X7R alone, but the coupling of both, in manner in which the opening of Panx1 leads to the formation of a large pore in P2X7R (Pelegrin and Surprenant 2009; Locovei et al. 2007; Iglesias et al. 2008).

To investigate pore formation by P2X7R in cultured rat cortical neurons, we used voltage-clamp technique and measure the responses to ATP, or its stable analogue BzATP. Figure 1a shows the typical responses of P2X7R, in accordance with previously reported findings (Arbeloa et al. 2012). Subsequently, to identify the involvement of Panx1 in the same neuronal culture, we measured the current–voltage relationship in the presence or absence of three Panx1 inhibitors carbenoxolone (100 μM), mefloquine (100 nM) and probenecid (1 mM). Figure 1b shows a substantial voltage-dependent contribution of Panx1 opening which is in agreement with previous findings (Pelegrin and Surprenant 2006; Iglesias et al. 2008) and at odds with other reports (Evans et al. 1996; Virginio et al. 1999). Discrepancies could be due to differences in the procedures and origin of neuronal cultures, or different Panx1 properties in the studied cells. Besides, we performed prolonged stimuli with ATP or BzATP, a non-hydrolysable and more efficient P2X7R agonist (Virginio et al. 1997), to induce the formation of large pores which took place only at a high concentration of BzATP (Fig. 2). These results are in contrast with other studies in which stimuli of short duration and lower concentration were sufficient to induce the formation of large pores. In parallel, we made similar stimuli with BzATP in cultured astrocytes and observe the typical behavior of the formation of a large pore (data not shown), just reported in other studies (Iglesias et al. 2009). This suggests that the formation of a large pore in cultured neurons depends not only on prolonged stimulation of P2X7R but also on concomitant membrane depolarization to induce Panx1 opening, as reported earlier (Locovei et al. 2007). Despite the differences found in several studies of the opening and closing dynamics, both Panx1 and P2X7R have been involved in the release of ATP into the extracellular medium in neurons or glial cells; however, the release mechanisms are still under study at present.

Fig. 1.

Ionic currents mediated by P2X7 receptors and Panx1 in cortical neurons. a Representative trace of ionic current induced by ATP and histogram showing the current amplitude in the presence or absence of BBG (50 nM), a P2X7 receptor antagonist. b Left, Current–voltage relationship (IV) evoked by voltage ramps of 10 s duration, ranging from −80 to 80 mV in the presence or absence of Panx1 inhibitors carbenoxolone (Cbx, 100 μM), probenecid (Prob, 1 mM) and mefloquine (MFQ, 100 nM). Right, histogram showing the current amplitude at +80 mV obtained from traces in IV relationship in the presence or absence of Panx1 inhibitors. Neuronal cultures and the external and internal solutions for whole-cell recordings were as described in a previous report (Arbeloa et al. 2012). Bars show the mean ± standard error. *P < 0.05 versus control (Ctrl), n > 10 for each bar

Fig. 2.

Effects of prolonged stimulation with ATP and BzATP in neurons. Left, representative whole-cell voltage-clamp current traces induced by prolonged stimulus with ATP or BzATP. The large gray bar above traces indicates the duration of stimulation. On the right, the average amplitude of the current is shown during (1) and after (2) applying ATP or BzATP. Extracellular bath solution with a pH of 7.3 contained the following (in mM): NaCl (140), KCl (5), CaCl2 (2), MgCl2 (1), HEPES (10), and glucose (10). The bars show the mean ± standard error. **P < 0.01 versus control (1), n > 7 for each bar

On the other hand, calcium homeostasis modulator 1 (CALHM1) is also involved in ATP release. CALHM1 was originally identified as a possible modifier of the age of onset of Alzheimer’s disease (Dreses-Werringloer et al. 2008; Lambert et al. 2010), and it is also able to form a large pore in mouse cortical neurons (Ma et al. 2012) as well as in the cells of the taste buds (Taruno et al. 2013). CALHM1 is a voltage and extracellular Ca²⁺-gated neuronal Ca²⁺-permeable ion channel that regulates cortical neuronal excitability in response to reduced extracellular Ca²⁺ concentrations (see Ma et al. 2012). In taste bud cells, CALHM1 is essential to detect sweet, bitter and umami flavors, and this process involves ATP release, which occurs also through CALHM1 (Ma et al. 2012). CALHM1 is structurally similar to connexins, innexins and pannexins, and although it has diverse functional activity, it forms a pore with a diameter similar to that estimated for connexins (Siebert et al. 2013). Thus, type II taste bud cells detect those flavors via G-protein-coupled taste receptors (Siebert et al. 2013) that stimulate a common signal transduction cascade involving activation of PLCB2, IP3-mediated Ca²⁺ release and Ca²⁺-dependent activation of TRPM5 channels (Zhang et al. 2003). This in turn depolarizes the plasma membrane to generate action potentials and subsequent non-vesicular release of ATP. CALHM1 is present in cerebro-cortical neurons whereby can also release ATP (Taruno et al. 2013). Figure 3 shows a hypothetical scenario where P2X7R, Panx1 and/or CALHM1, may be involved in release of ATP in neurons or glial cells.

Fig. 3.

Simplified hypothetical scenario of ATP release mechanisms in brain cells through channels or receptors. Physiological membrane depolarization (ΔVm) may activate Panx1 or CALHM1 in a voltage-dependent manner. In turn, the release of ATP through Panx1 or CALHM1 activates P2X7R which induces influx of Ca²⁺ and a concomitant local extracellular decrease in the concentration of this cation, which may activate CALHM1. On the other hand, prolonged activation of P2X7R with ATP (prolonged ATP) may induce pore formation and further ATP release. In normal conditions, ion concentrations may return to normal by operation of ATP-dependent pumps or by activation of rectifying channels. The return at normal extracellular [Ca²⁺] induces inactivation of Panx1 and CALHM1. During ischemia ion concentrations are altered as a consequence of energy failure and miss-operation of ATP-dependent pumps and may cause activation of Panx1 and CALHM1, release of ATP and sustained activation of P2X7R, which results in intracellular Ca²⁺ elevation and excitotoxicity

ATP Excitotoxicity: The Case of Cerebral Ischemia

Excessive release of ATP can cause cell death (Matute et al. 2007; Arbeloa et al. 2012). Although little is known about the relevance of ATP damage as a primary event in the pathophysiology of acute and chronic diseases, the release of ATP from dying cells may well contribute as a secondary mechanism to aggravate the extent of ongoing damage in numerous pathological conditions (Matute et al. 2007). Thus, during and after stressful events and in chronic and acute neurological disorders such as ischemia, damaged neurons and glial cells release ATP into the extracellular space (Braun et al. 1998; Juranyi et al. 1999; Melani et al. 2005; Rossi et al. 2007; Yenari et al. 2010).

ATP in excess may over activate P2X7R and trigger signaling cascades leading to neurodegeneration after ischemia (Le Feuvre et al. 2003) and traumatic damage to the spinal cord (Wang et al. 2004). In particular, the expression of P2X7R is upregulated in experimental transient cerebral ischemia (Cavaliere et al. 2004; Sperlágh et al. 2006), and higher levels of P2X7R can be secondary mediators of damage (Le Feuvre et al. 2003). Furthermore, we recently found that P2X7R participates in ATP excitotoxicity in neurons during ischemia, as a consequence of cytosolic Ca²⁺ overload (Arbeloa et al. 2012). In turn, other studies also revealed the involvement of Panx1 in cerebral ischemia. Thus, brain energy deprivation causes anoxic depolarization which, if prolonged, leads to irreversible neuronal and glial death (Somjen 2001; Lipton 1999). Panx1 is activated during ischemia in hippocampal and cortical neurons along with large post-anoxic ion currents which were partly inhibited by Panx1 blockers, indicating that Panx1 is a significant component of ischemia outcome (Thompson et al. 2006). Furthermore, Panx1 is localized near NMDA receptors in the postsynaptic density (Zoidl et al. 2007), thus, it can be activated during synaptic activity. Additionally, the identification of Panx1 as a component of the large post-anoxic ion currents suggests that Panx1 is a relevant player in neurotransmitter excitotoxicity and in ischemic cell death.

In addition to P2X7R or Panx1, other intermediaries may contribute to the massive release of ATP during the early stages of ischemia. Indeed, activation of P2X7R promotes the opening of Panx1 (Locovei et al. 2007) through an intermediate route mediated by phosphorylation (Iglesias et al. 2008). Moreover, the high-permeability pore formed by prolonged P2X7R activation does not occur through pannexins and suggests that other signaling pathways may be involved (Alberto et al. 2013). Alternatively, other studies show that in macrophages the activation of Panx1 can be mediated by P2X7R through a protein–protein interaction (Pelegrin and Surprenant 2006). Moreover, maxi-anion channels release ATP under ischemic conditions in astrocytes (Liu et al. 2008; see Sáez and Leybaert 2014). Finally, another possible player, CALHM1, is a putative contributor to ATP release (Ma et al. 2012; Taruno et al. 2013) which in turn makes it a candidate to contribute to excitotoxicity by ATP.

Conclusions

Like glutamate, extracellular ATP is a major excitatory neurotransmitter in the CNS via activation of ionotropic P2X receptors. The mechanisms by which intracellular ATP is released are a matter of intense debate, and there are data supporting the involvement of P2X7R as well as of Panx1. In addition, new evidence also suggests a possible role of CALHM1 in ATP release. We hypothesize that these three channels are major contributors to physiological ATP release (see Fig. 3); however, it is possible that anionic maxi-channels and other as yet unknown mechanisms may also make a substantial contribution to non-vesicular ATP release. On the other hand, massive release of ATP takes place in neuroinflammation associated with neurodegenerative diseases, as well as in ischemic and traumatic brain injuries. ATP released from neurons and/or glial cells, such as that taken place in ischemia, may be mediated by P2X7R or Panx1, whereas the contribution of CALHM1 remains to be elucidated. Thus, prolonged activation of P2X7R may promote the formation of a large pore in neural cells, efflux of ions and ATP itself, and subsequent irreversible depolarization which leads to cell death. Pore formation may involve P2X7R but not Panx1, or vice versa, thus, drugs inhibiting these channels may offer therapeutic potential in ischemia and brain diseases in which ATP-mediated excitotoxicity contributes to propagate cell damage.